The first session on this first full day at ECREA 2018 is on polarisation, and starts with Melanie Magin. She begins by highlighting the potential deleterious effects of polarisation on society: societies need a common meeting ground, and this has traditionally been provided by the news media and their agenda-setting function. But the diversification of information sources and channels may contribute to fragmenting this, and the algorithmic selection of content in these channels could aid this fragmentation – yet there is very little empirical evidence for the existence of the echo chambers or filter bubbles this is said to cause.
Echo chambers and filter bubbles could also simply be a niche phenomenon that mainly affects politically extreme users; overall, then, there is a need to develop new indicators that assess the diversity of issues that individuals consider to be important for society; that identify the top issue that individuals highlight in contemporary society; and that assess the overlap in key issues between any set of individuals. The more compatible these issue horizons are between individuals, the less fragmented is society.
Such issue horizon compatibility is also affected by news sources. Legacy, generalist news media should lead to greater issue compatibility; personalised, algorithmically curated social media should produce lower levels of issue compatibility. These effects may also be further moderated by the strength and extremity of users’ political attitudes.
The project tested these metrics using two-week daily online diaries with some 333 German Internet users; of these, some 57% mentioned the refugee crisis as their top issue, and individuals nominated up to 28 issues over the course of the two weeks. News media appeared to have a positive effect on the compatibility of issue horizons, but it turned out that they made individual issue horizons more diverse especially for more politically extreme users. Social media turned out not to have any significant effect. Using news media also made issue horizons more compatible across users, but against only for users at the political extremes.
So: news media do not increase the compatibility of issue horizons; social media do not decrease the compatibility of issue horizons. The fears of negative effects of algorithmically based information sources on societal cohesion is clearly exaggerated, but at the same time such social media also do not have any positive effects either. Mainly, it appears to be the news media that make issue horizons more compatible amounts those with extreme political attitudes. However, the fact that this was a German study – where social media play a fairly limited role – should also be recognised here.